Despite recent improvements in our understanding of the biology of breast cancer, few recent advances have been made in the systemic therapy of this disease, and present statistics indicate that 3.5% of American women will die of breast cancer. Initial treatment of advanced breast cancer is based on the hormone-receptor status of the tumor determined by in vitro assay for the ER in tumor tissue. Patients with advanced hormone- sensitive (ER+) cancer are treated by hormonal manipulation, and 55-60% of these women will respond favorably. Additional information is needed to predict response in individual patients more accurately, based on an in vivo functional assessment of estrogen receptor (ER) status and assessment of other factors predictive of tumor aggression; many such prognostic factors, including several nuclear proteins, have been evaluated, but none has currently proven more accurate than the status of the axillary nodes at diagnosis. In patients with metastatic ER+ disease who are treated with hormonal therapy, 7-10% will experience a symptomatic "flare reaction". This reaction has been attributed to the temporary agonist effect of the hormone and cannot be distinguished from disease progression clinically or radiologically, but is associated with subsequent response to that therapy. We have previously shown that positron emission tomography (PET) with 16alpha-[18F]fluoro-17beta-estradiol (FES) provides reliable in vivo information about the ER status of the tumor. Quantitative results of PET with [18F]-2-fluoro-2 deoxyglucose (FDG) have been shown by others to predict grade and behavior of several types of malignant tumors including breast cancer. We propose to perform serial PET with both FES and FDG and serial tumor biopsies (for assessment of several immunohistochemical prognostic factors) in women with locally advanced ER+ breast cancer before and early after initiation of tamoxifen therapy. It is expected that the combination of functional data provided by PET and that provided by prognostic-factor assessment will yield information about tumor aggression and hormonal sensitivity so that hormone therapy of advanced breast cancer can be individualized. We believe that a subclinical flare detected by PET imaging early after initiation of treatment may predict responsiveness to hormone therapy in this group of patients. Additionally, the results of this study should be translatable to patients with metastatic breast cancer and, hence, provide a method for distinguishing between the clinical flare reaction and disease progression.